MODEL REGRESI LINIER BERGANDA MENGGUNAKAN PENAKSIR PARAMETER REGRESI ROBUST M-ESTIMATOR (Studi Kasus: Produksi Padi di Provinsi Jawa Barat Tahun 2009)


Rini Cahyandari(1*), Nurul Hisani(2)

(1) Jurusan Matematika Fakultas Sains dan Teknologi Universitas Islam Negeri Sunan Gunung Djati Bandung, Indonesia
(2) Jurusan Matematika Fakultas Sains dan Teknologi Universitas Islam Negeri Sunan Gunung Djati Bandung, Indonesia
(*) Corresponding Author

Abstract


The least squares method is one method of parameter estimation in regression models. This method produces an unbiased estimator with consideration the assumptions are linearity, non-multicollinearity, non-autocorrelation, homoscedastic, and normally distributed error fulfilled. When the assumptions are not fulfilled, such as error distribution is not normal due to the existence of outliers, an estimation obtained are not exact.An alternative method that can overcome the problem of outliers is robust regression method using M-estimator (Iteratively Reweighted Least Squares). M-estimator is an iterative method using weighting function Huber and Tukey bisquare to estimate the parameters or coefficients in the regression model. The best model obtained by M-estimator robust method using Huber and Tukey bisquare is determined by the value R2adjusted and standard error values.

Full Text:

PDF

References


Badan Pusat Statistika. 2010. Jabar Dalam Angka 2010.

Chen, Colin. 2002. Robust Regression and Outlier Detection with the ROBUSTREG Procedure. Paper 265-267. SAS Institute: Cary, NC

Fox, John. 2002. Robuts Regression Appendix to An R and S-PLUS Companion to Applied Regression.

Gujarati, Damodar N. 2004. Basic Econometrics. The McGraw Hill-Companies.

Hisani, Nurul. 2012. Model Regresi Linier Berganda Menggunakan Penaksir Parameter Regresi Robust M-estimator (Iteratively Reweihgted Least Squares) Melalui Software R (Studi Kasus: Produksi Padi di Provinsi Jawa Barat Tahun 2009). Bandung: Skirpsi S1 Matematika UIN SGD.

J. Rousseeuw, Peter and M. Leroy Annick. 1987. Robust Regression and Outlier Detection. Canada: John Wiley & Sons Inc.

Norman R. Draper, Harry Smith. 1998. Applied Regression Analysis. USA: John Wiley & Sons Inc.